Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/tuvimen/reliq

HTML parsing and searching tool
https://github.com/tuvimen/reliq

c html parsing searching

Last synced: 23 days ago
JSON representation

HTML parsing and searching tool

Awesome Lists containing this project

README 

        
# reliq



reliq is a html parsing and searching tool.



## Building



### Build



    make install



### Build library



    make lib-install



### Build linked



    make linked



### AUR



    yay -S reliq



## Usage



Get some help



    man reliq



Get `div` tags with class `tile`.



    reliq 'div class="tile"'



Get `div` tags with class `tile` as a word and id `current` as a word.



    reliq 'div .tile #current' index.html



Get tags which does not have any tags inside them from file `index.html`.



    reliq '* c@[0]' index.html



Get empty tags from file 'index.html'.



    reliq '* m@>[0]' index.html



Get hyperlinks from level greater or equal to 6 from file `index.html`.



    reliq 'a href @l[6:] | "%(href)v\n"' index.html



Get any tag with class `cont` and without id starting with `img-`.



    reliq '* .cont -#b>img-' index.html



Get hyperlinks ending with `/[0-9]+.html`.



    reliq 'a href=Ee>/[0-9]+\.html | "%(href)a\n"' index.html



Get links that are not at 1 level, both are equivalent.



    reliq 'a href l@[!0] | "(href)v\n"' index.html

    reliq 'a href -l@[0] | "(href)v\n"' index.html



Get `li` tag that does not start with `Views:`, both are equivalent.



    reliq 'li m@bv>"Views:"' index.html

    reliq 'li -m@b>"Views:"' index.html



Get `ul` tags and html inside `i` tags that are inside `p` tags.



    reliq 'ul, p; i | "%i\n"' index.html



Get the second to last image in every `li` with `id` matching extended regex `img-[0-9]+`.



    reliq 'li #E>img-[0-9]+; img src [-1] | "%(src)v\n"' index.html



Get `tr` and `td` inside `table` tag.



    reliq 'table; { tr, td }' index.html



Get `tr` with either `title` attribute with class `x1` or `x2`, or if it has one child. Note that `)` has to be preceded with space otherwise it will be a part of previous  argument and that brackets have to touch for it to be or.



    reliq 'tr ( title ( .x1 )( .x2 ) )( c@[1] )'



Process output using `cut` for each tag, and with `sed` and `tr` for the whole output.



    reliq 'div #B>msg_[0-9-]* | "%(id)v" cut [2] "-" / sed "s/^msg_//" tr "\n" "\t"' index.html



Process output in a block (note that things in blocks have to be separated by `,` as just newline is not enough).



    reliq '

        {

            div class=B>"memberProfileBanner memberTooltip-header.*" style=a>"url(" | "%(style)v" / sed "s#.*url(##;s#^//#https://#;s/?.*//;p;q" "n",

            img src | "%(src)v" / sed "s/?.*//; q"

        } / sed "s#^//#https://#"

    ' index.html



Get `tr` in `table` with level relative to `table` equal `1`, and process it individually for every tag in block, and at the end of each block delete every `\n` character and append `\n` at the end.



    reliq '

        table border=1; tr l@[1]; {

            td; * c@[0] | "%i\t"

        } | tr "\n" echo "" "\n"

    ' index.html



Same but process all tags at once in block, then process final output of the block deleting all `\n` and appending `\n` at the end. The above creates tsv where each `tr` has its own line, but this example craetes only one line.



    reliq '

        table border=1; tr; {

            td; * c@[0] | "%i\t"

        } / tr "\n" echo "" "\n"

    ' index.html



### JSON like output



Output of expression can be assigned to a field. Fields combined in json like structure will be written at the end of output, if any expression is not assigned to a field it's output will be written before the structure.



    reliq '.links.a a href | "%(href)v\n", img src | "%(src)v\n"'



will return



```

/static/images/icons/wikipedia.png

/static/images/mobile/copyright/wikipedia-wordmark-en.svg

/static/images/mobile/copyright/wikipedia-tagline-en.svg

//upload.wikimedia.org/wikipedia/commons/thumb/0/08/Sodium_nitrite.svg/220px-Sodium_nitrite.svg.png

https//upload.wikimedia.org/wikipedia/commons/thumb/a/a6/Nitrite-3D-vdW.png/110px-Nitrite-3D-vdW.png

https//upload.wikimedia.org/wikipedia/commons/thumb/0/05/Sodium-3D.png/80px-Sodium-3D.png

{"links":["/wiki/Main_Page","/wiki/Wikipedia:Contents","/wiki/Portal:Current_events","/wiki/Special:Random","/wiki/Wikipedia:About"]}

```



Note that from now on any json structure in examples will be prettified for ease of reading, in reality reliq returns compressed json.



It is a json like structure because reliq doesn't enforce json output as in above example, if output would be directly connected to json parser an error might accur when incorrect changes are made to reliq script.



It also does not check for repeating field names e.g.



    reliq '.a span #views | "%i", .a time datetime | "%(datetime)v\a"'



will return incorrect json



```

{

  "a": "29423",

  "a": "2024-07-19 07:12:05"

}

```



So reliq returning json depends on user!



Field is specified when there's `.` followed by string matching `[A-Za-z0-9_-]+` before expression e.g.



    reliq '.some_fiELD-1 p | "%i"'



Fields can only be specified before expression, and will be ignored inside it making incorrect patterns e.g.



    reliq 'ul; .a li' index.html



This will not return json like structure, or anything since `.a` was specified as plain name of tag which cannot exist in html.



Fields take precedence before everything



    reliq '

        .field dd; {

            time datetime | "%(datetime)v\a",

            a m@v>"<" | "%i\a",

            * l@[0] | "%i"

        } / tr '\n' sed "s/^\a*//;s/\a.*//"

    ' index.html



Fields can be nested in another fields



    reliq '

        .user {

            h4 class=b>message-name,

            * .MessageCard__user-info__name

        }; {

            .name [0] * c@[0] | "%i",

            .link [0] a class href | "%(href)v",

            .id.u [0] * data-user-id | "%(data-user-id)v"

        },

    ' index.html



will return



```

{

  "user": {

    "name": "TUVIMEN",

    "link": "https://examplesite.com/u/TUVIMEN/",

    "id": 1245

  }

}

```



If field that nests other fields has expression without field, that expression will not be assigned to nesting field, and will output normally before json structure, breaking compability with json e.g.



    reliq '

        .user div #user; {

            .name span .user-name | "%i",

            .link [0] a href | "%(href)v",

            div .user-avatar; img src | "%(src)v\n"

        }

    ' index.html



will return



```

https://exampleother.com/a/8122.jpg

{

    "user": {

        "name": "hexderm",

        "link": "https://exampleother.com/users/8122"

    }

}

```



This further underlines the importance of writing correct script.



Unless fields are nested they will be on the same level no matter where they are in script



    reliq '

        .signature div .signature #B>sig[0-9]* | "%i",

        dl .postprofile #B>profile[0-9]*; {

            dt l@[1]; {

                .avatar img src | "%(src)v",

                .user a c@[0] | "%i",

                .userid.u a href c@[0] | "%(href)v" / sed "s/.*[&;]u=([0-9]+).*/\1/" "E",

            },

            .userinfo.a("\a") dd l@[1] m@vf>" " | "%i\a" / tr '\n\t' sed "

                s/([^<]*)<\/strong>/\1/g

                s/ +:/:/

                /